March 28, 2022 PAO-03-022-CL-24
Pharmaceuticals and medical devices require both primary and secondary packaging, both of which often include critical information. Drug products are packaged in vials, syringes, IV bags, blister packs, bottles, and so on. Those containers are then placed into cartons, sleeves, and pallets. A final product will likely have a label on each primary and secondary container.
Those different containers comprise a wide variety of materials, including glass, plastic, metal foil, paper, and cardboard. Some are rigid, while others are flexible. The materials may be glossy, matte, or transparent. Some labels come preprinted, and others are printed after the label is applied to the container. Some materials — glass, acrylics, plastic foils — are influenced by light. Each of these different types of surfaces may affect the way applied print appears and is perceived.
Regardless of the substrates and their surface properties, labels must be inspected according to regulatory requirements to ensure accuracy. For instance, on flexible packaging, lettering often does not look exactly like it would on a flat surface — it will be slightly distorted. For glossy and transparent surfaces, any ambient light can change the appearance of a label. Automated inspection systems must be programmed to cope with these variations in order to provide robust and reliable visible inspection of printed pharmaceutical labels.
Different areas of text on a given label have different levels of criticality. There are regulatory guidelines, which may differ from region to region. For products going to multiple markets, therefore, it is necessary to comply with multiple standards. Typical standards include GAMP (Good Automated Manufacturing Practice) guidelines and ISO (International Standards Organization) standards.
On the basis of these guidelines and standards, drug makers are responsible for classifying which areas of their printed labels are highly critical. Typically, those areas include the chemical makeup and the dosage frequency and level. Small errors in these areas can be quite critical. A missing full stop (period) in a dosage quantity could change the dosage amount by a factor of 10. Similarly, missing the dash in a dosage frequency would change a range of a few hours (e.g., 2–4 hours) to a much longer period time (24 hours). In these cases, a small change would have a tremendous impact and is absolutely critical to catch. A less important error would be a full stop missing in the address of the manufacturing company.
IV-bag close-up. Image courtesy of Vitronic.
Visual inspection systems have to be sufficiently flexible to assess the many different types of labels attached to many substrates and intended to meet varying regulatory requirements with an ability to define critical inspection areas that will also vary from product to product.
Automated inspection systems offer several improvements over visual inspection by human operators. One of the primary advantages is the ability to inspect 100% of the information on manufactured drug products. Human inspectors can get tired and distracted and focus less over time on the details. There are also variations in performance from operator to operator. In addition, operator-based inspections are generally performed at set periods on selected sample products based on the assumption that the products evaluated will be representative of all of the products being produced, rather than inspecting all products.
Automated systems can function 100% of the time and are able to inspect every single piece of information on the primary and secondary packaging. Operators are no longer needed for this activity and are freed up to perform more value-generating activities. Inspection machines can also be programmed to check labels in multiple languages, something most operators are not capable of doing. Furthermore, automated inspection results are more consistent, because operator variability is eliminated. They can also be programmed to prioritize different areas of interest with respect to criticality.
Trending of results is another important benefit of using automated inspection systems. Print placement and quality, for instance, can be monitored for drift. The software can also be programmed to report when certain error levels occur. Importantly, the inspection system can interface with other operating systems (e.g., SAP, MES, LMS), enabling analysis of the data on an ongoing basis and allowing drug manufacturers to see how their processes are performing. Specific data are also available to analyze with an automated system; unlike when an operator pulls a unit off the line, when an automated system fails a label, there are stored images that can be viewed to determine the cause, which also provides greater insight into performance. All of this information allows for greater process understanding and the ability to take actions to address issue before they become real problems, as well as to pursue continuous improvement initiatives.
In addition, automated inspection systems can include built-in centralized password control and auditing capabilities. Any changes typically require validation by a second person to ensure that they are correct. The ability to track changes, including what was done when and by whom, can be extremely valuable for determining root causes if products with printing errors are assessed as meeting specifications.
Depending on the size of the container and the complexity of the print data, automated inspections are completed by one machine in the range of one per every two seconds (IV bags) to ten per second (vials). Systems are scalable, however, simply by adding additional cameras and running them in parallel. With this approach, it is possible to match the rate of inspection with that of product production.
To function effectively, automated visual printing inspection systems must be able to acquire accurate images of the product label being evaluated. That requires use of appropriate camera and lighting technologies, the latter of which are typically LED-based today. The lighting system is designed to eliminate any ambient lighting effects.
These physical aspects of the system must be combined with effective software that allows comparison of data from the label image to the data in the initial input file (e.g., PDF or Word document,) describing what the label should look like. Proprietary algorithms are used to take account of distortions for flexible packaging, reflections for glossy surfaces, and other considerations. These algorithms allow for accurate determination of any differences between the print image and the desired appearance.
Sample teach-in process. Video courtesy of Vitronic.
During the automated teach-in process, the tolerance class is precisely defined in each text field. Image courtesy of Vitronic.
Beyond the ability of automated visual inspection services to do a more objective job of visual inspection than humans can and with greater productivity, they also enable the inspection of features that a human simply cannot do. One example is two-dimensional barcodes, which would be too complex for humans to inspect but not a problem for machine vision. Another is anti-counterfeiting technology incorporated into packaging.
Counterfeiting is an increasingly significant issue, and automated visual inspection is enabling drug manufacturers to implement anticounterfeiting measures that can have a significant impact on increasing the safety and cost-effectiveness of the supply chain. The most common anti-counterfeiting elements incorporated into labels or packaging are unique device identifiers (UDIs). They make it possible for drug manufacturers to track and trace from the origin of the material supply of the product through the manufacturing plant and distribution channels to hospitals and clinics and ultimately patients.
UDIs, however, comprise not just text but often data matrix code and can be very small on small drug product packaging. As a result, they are not effectively inspectable by the human eye; robust automated vision systems are required to properly read the codes. The dramatic growth in numbers of product units accompanying the rising prevalence of personalized medicines produced in smaller batch sizes serves as a further driver for the use of automated inspection systems.
Vitronic is focused solely on the development of machine vision systems. In addition to the pharmaceutical industry, we develop solutions with applications in traffic control and public safety, logistics, the automotive industry, photovoltaics, and human body measurement. The knowledge gained from working in these other fields can often be leveraged in the design of systems for medical device and pharmaceutical packaging inspection.
As one example, technology we have developed for detecting and tracking surface defects in the solar industry has been used in our visual inspection systems for assessing pharmaceutical labels for defects..
In addition, within the pharmaceutical industry, automated inspection systems are used for various applications beyond printed label review. For parenteral products packaged in glass vials, for instance, the fill level must be confirmed, the glass evaluated for any cracks, the drug product inspected for particulates, and the container closure for proper positioning and sealing.
Vitronic has a different setup for automated label inspections compared with automated vial inspections, since the camera and lighting technologies employed are optimized for the specific application. However, the main differences lie in the software for failure detection and the fact that labels are planar, while determining particulates in vials requires three-dimensional analysis. Our product development software and hardware teams leverage real-world data to develop new tools for the future, and, as more is learned about the performance of our systems in different applications within the pharmaceutical industry, that knowledge is applied wherever it can improve performance.
Vitronic is invested in the development of high-performance automated vision systems that can help our pharmaceutical customers increase their productivity while ensuring the highest product quality. Our success can only be achieved if our customers consistently deliver high-quality products to patients. For that reason, Vitronic seeks to establish long-term, consultative relationships with our clients. We work very closely with each and every customer to design systems tailored to their specific products and processes –– not only the products they are working with today but those they might be manufacturing in the future.
Ideally, a customer will involve Vitronic at the product development stage to ensure the design of product packaging best suited for automated visual inspection. We can provide advice on what elements will be easy or difficult to inspect and how changes to the design will make it easier for our system to detect defects.
Even after a system has been installed, we stay in regular communication with our clients to ensure the continued performance of our vision inspection solutions. We often receive images of failures from production lines along with requests for assistance in how to address them. It is also important for us to remain aware of changes at the customer that might impact that performance and to share new developments at Vitronic that might benefit them.
Florian joined VITRONIC’s healthcare division in 2018. He is responsible for automated visual inspection solutions for pharmaceuticals and medical devices. The main focus of his work as a Product Manager is the identification of process requirements in difficult to inspect products and translating into new standardized solutions. Before joining VITRONIC Florian gained experience in medical device, hygiene and environmental technology industries. He holds an MBA and an engineering degree in Technical Health Care.